Re: from 415k to under 12k, from a random input

On Apr 14, 7:55 pm, SG <s.gesem...@xxxxxxxxx> wrote:
On 14 Apr., 21:41, Industrial One <industrial_...@xxxxxxxxxxx> wrote:



On Apr 14, 6:33 pm, SG <s.gesem...@xxxxxxxxx> wrote:
There's a difference between "every file in the whole wide world" and
"every possible file". You probably don't even need more than 500
bits to enumerate all files from the whole wide world (including
future and past files). The problem is, however, the nontrivial &
injective mapping from the set of all files from the whole wide world
to the set of {0,1}^500 and the mapping's inverse.

I encourage you to find an efficient implementation of a mapping like
this and to compress everything to 500 bits. This would be awesome!

You got it all wrong. That's like saying you can compress every movie
in the world to 20 bits because there are less than a million existing
movies. You have to be able to predict the bit-by-bit contents of the
next movie to compress it down to that much. Say every existing Mickey
Mouse film has him with 2 years, and there are imaginary films with
Mickey mouse having only one/no ears that DON'T exist. But how do you
KNOW the next Mickey Mouse movie won't draw him without ears? Because
that bit exists out of the 20 bits of existing movies and is part of
the 2^1000000000000000000 that don't exist yet. Get it?

I was being hypothetical and sarcastic.
Get it?

Cheers!
SG

Yeah, I was just contributing to JG's education and thought Mickey
Mouse would get his attention.
.

Relevant Pages
Loading